Recent searches

No recent searches

Search options

Only available when logged in.
techhub.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A hub primarily for passionate technologists, but everyone is welcome

Administered by:

Server stats:

4.6K
active users

techhub.social: AboutStatusProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.4.3

#parameters

1 post1 participant0 posts today
Karsten Schmidt

#ReleaseMonday — New version (v0.27.0) of thi.ng/genart-api, a platform-independent extensible API for browser-based computational/algorithmic/generative art projects:

This version features an overhaul of the platform provided PRNG (pseudo-random number generator) handling and makes it easier to create multiple PRNGs for artworks which require/desire them...

Related section in the README:
github.com/thi-ng/genart-api/b

Also, just as a reminder, the project has:

- no external dependencies
- adapters for 3 art platforms (EditArt, fxhash, Layer)
- 6 example projects
- testing/dev sandbox with two parameter editors
- WebAssembly bindings & demo (currently for #Zig only)

Happy coding! :)

thi.ng/genart-apithi.ng/genart-api
#GenArtAPI#GenerativeArt#AlgorithmicArt
N-gated Hacker News

Ah, behold the majestic #DeepSeekR1-0528, a model so #mysterious and elusive that not even #Inference #Providers dare to touch it. 🤔✨ With a grand total of zero downloads last month, it's clear that this #685B parameter behemoth is the hottest #AI sensation—if only in its creator's wildest dreams. 🐒💭
huggingface.co/deepseek-ai/Dee #Parameters #HottestSensation #HackerNews #ngated

huggingface.codeepseek-ai/DeepSeek-R1-0528 · Hugging FaceWe’re on a journey to advance and democratize artificial intelligence through open source and open science.
💧🌏 Greg Cocks

Don’t Move A Mussel - The Role Of Key Environmental Drivers And Management Scale In Assessing Spatial Variation In Dreissenid Spread Risk In The Missouri River Basin
--
doi.org/10.1016/j.ecolind.2024 <-- shared paper (Missouri River Basin)
--
usgs.gov/centers/western-fishe <-- shared technical article (Columbia River Basin)
--
rconnect.usgs.gov/webDR/ <-- shared link to webDR: Dreissenid Risk (Missouri River Basin)
--

@USGS

N-gated Hacker News

Ah, the age-old debate: should we pass arguments individually and make our #code as readable as ancient hieroglyphs 🗿, or shove them into a single #object and pretend we’re sophisticated? 🤔 Spoiler alert: object #parameters are the holy grail ✨, and if you disagree, you’re clearly living in 2025 BC, not 2025 AD.
carlos-menezes.com/single-para #readability #programming #debate #software #development #HackerNews #ngated

www.carlos-menezes.comn-params vs single param - Carlos Menezes
arXiv cs.CV bot

Towards Efficient Partially Relevant Video Retrieval with Active Moment Discovering

Peipei Song, Long Zhang, Long Lan, Weidong Chen, Dan Guo, Xun Yang, Meng Wang
arxiv.org/abs/2504.10920 arxiv.org/pdf/2504.10920 arxiv.org/html/2504.10920

arXiv:2504.10920v1 Announce Type: new
Abstract: Partially relevant video retrieval (PRVR) is a practical yet challenging task in text-to-video retrieval, where videos are untrimmed and contain much background content. The pursuit here is of both effective and efficient solutions to capture the partial correspondence between text queries and untrimmed videos. Existing PRVR methods, which typically focus on modeling multi-scale clip representations, however, suffer from content independence and information redundancy, impairing retrieval performance. To overcome these limitations, we propose a simple yet effective approach with active moment discovering (AMDNet). We are committed to discovering video moments that are semantically consistent with their queries. By using learnable span anchors to capture distinct moments and applying masked multi-moment attention to emphasize salient moments while suppressing redundant backgrounds, we achieve more compact and informative video representations. To further enhance moment modeling, we introduce a moment diversity loss to encourage different moments of distinct regions and a moment relevance loss to promote semantically query-relevant moments, which cooperate with a partially relevant retrieval loss for end-to-end optimization. Extensive experiments on two large-scale video datasets (\ie, TVR and ActivityNet Captions) demonstrate the superiority and efficiency of our AMDNet. In particular, AMDNet is about 15.5 times smaller (\#parameters) while 6.0 points higher (SumR) than the up-to-date method GMMFormer on TVR.

arXiv logo
arXiv.orgTowards Efficient Partially Relevant Video Retrieval with Active Moment DiscoveringPartially relevant video retrieval (PRVR) is a practical yet challenging task in text-to-video retrieval, where videos are untrimmed and contain much background content. The pursuit here is of both effective and efficient solutions to capture the partial correspondence between text queries and untrimmed videos. Existing PRVR methods, which typically focus on modeling multi-scale clip representations, however, suffer from content independence and information redundancy, impairing retrieval performance. To overcome these limitations, we propose a simple yet effective approach with active moment discovering (AMDNet). We are committed to discovering video moments that are semantically consistent with their queries. By using learnable span anchors to capture distinct moments and applying masked multi-moment attention to emphasize salient moments while suppressing redundant backgrounds, we achieve more compact and informative video representations. To further enhance moment modeling, we introduce a moment diversity loss to encourage different moments of distinct regions and a moment relevance loss to promote semantically query-relevant moments, which cooperate with a partially relevant retrieval loss for end-to-end optimization. Extensive experiments on two large-scale video datasets (\ie, TVR and ActivityNet Captions) demonstrate the superiority and efficiency of our AMDNet. In particular, AMDNet is about 15.5 times smaller (\#parameters) while 6.0 points higher (SumR) than the up-to-date method GMMFormer on TVR.
💧🌏 Greg Cocks

A Site Selection Framework For Urban Power Substation At Micro-Scale Using Spatial Optimization Strategy And Geospatial Big Data
--
doi.org/10.1111/tgis.13093 <-- shared paper
--
“In this study, [they] model spatiotemporal heterogeneity and incorporate it into optimizing the location of substations. The optimized substation placement ensures electrical service coverage for over 99% of the area during peak power usage seasons, compared to the current coverage of 72%...”

Replied in thread
*
Radio Azureus

@dazo
Thank you for that valuable input

I've looked further into it and saw that it is quite extensive to set up and also has a suit that seems to be more targeted at larger environments to back up in a nice and cool manner, double confirming your words

When I discovered that for me it was easier to just jump back to bacula backup which I have known for decades

bacula.org/

#bash#csh#ksh
Continued thread
*
Dendrobatus Azureus

You can download a live image of gparted and work with the latest version with ease. My debian based distro has GParted 1.3.1 which is quite behind v1.7.0-1

Im downloading the latest right now!

log
$ wget -c cfhcable.dl.sourceforge.net/pr
--2025-03-16 11:54:11-- cfhcable.dl.sourceforge.net/pr
Resolving cfhcable.dl.sourceforge.net (cfhcable.dl.sourceforge.net)... 146.71.73.5
Connecting to cfhcable.dl.sourceforge.net (cfhcable.dl.sourceforge.net)|146.71.73.5|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 589299712 (562M) [application/octet-stream]
Saving to: ‘gparted-live-1.7.0-1-amd64.iso?viasf=1’

gparted-live-1.7.0-1 9%[==> ] 55.49M 286KB/s eta 28m 15s
^Z

gparted.org/livecd.php

#gparted#partitions#bash
*
Dendrobatus Azureus

Giving credit to the programmers of GPARTED(8)

gparted works its magic, by entering correct parameters to a suite of partition control & editing commands, which are sh envoked, so you can easily manipulate your partitions on all your SSDs HDDs from the comfort of your UI

When you want to batch manipulate partitions, you can study the log output and make sh scripts yourself, controlling partitions anywhere.
You also have the convenience of running gparted from sh so it still works its magic for you, without the UI!

I usually run cfdisk gdisk fdisk when I partition a fresh mechanical or SSD, later on I invoke gparted when I want to resize or move them

it also runs important commands at the end so that the kernel gets to know your new partition layout, which makes rebooting your machine to use them unneeded

I shrunk and resized a partition where I installed a program, which needed 75GB (*1024!) as installation space but only uses 56GB in the end. I left 12GB of breathing room on the partition after the shrink and of course grew the partition before with the same size, minus the alignment snip of 1MB

log:
myserver kernel: JFS: nTxBlock = 8192, nTxLock = 65536
myserver kernel: SGI XFS with ACLs, security attributes, realtime, scrub, repair, quota, debug enabled
myserver kernel: sdb: sdb1 sdb2 sdb3 sdb4
myserver kernel: sdb: sdb1 sdb2 sdb3 sdb4
^Z

@altbot

gparted.org

#gparted#partitions#bash
TrendingLive feeds
About